AI brings perovskite solar cells closer to sustainable commercialization

What if we could get clean electricity more cheaply and protect the environment at the same time? A team of researchers in Korea has presented a roadmap for sustainable solar cells that integrates AI technology to bring this future closer. Their study is published in Green Chemistry and was selected as the cover of the online edition of the journal.

Solar power is a prominent sustainable energy source that can reduce carbon emissions. In particular, perovskite solar cells, also known as “next-generation solar cells,” have received attention for their high theoretical efficiency (34%), which surpasses that of traditional silicon solar cells. However, the production of these cells requires the use of toxic chemicals and has limited long-term stability, hindering their commercialization.

POSTECH (Pohang University of Science and Technology) Department of Chemical Engineering, led by Professor Jeehoon Han, collaborated with the Department of Chemical Engineering at University of Seoul, led by Professor Min Kim, to solve the problem. The research team developed a new manufacturing process that uses bio-based solvents, such as gamma-valerolactone (GVL) and ethyl acetate (EA), instead of the toxic solvent dimethylformamide (DMF).

The study’s core is AI-based reverse engineering technology. By analyzing experimental data, the research team identified optimal conditions to boost efficiency while minimizing costs and carbon emissions. They then verified the conditions proposed by AI through actual experiments and presented a sustainability evaluation model that considers manufacturing costs, environmental impact, and process efficiency, as well as a global deployment scenario.

The GVL-EA process developed by the research team reduces the manufacturing cost of perovskite solar cells by half and decreases the climate impact by over 80%. Additionally, they confirmed that considering module lifespan and recycling strategies together can help identify the actual break-even point for commercialization in different regions.

POSTECH’s Professor Han said, “AI has found conditions that were previously considered impossible by optimizing the process itself.” He added that using non-toxic bio-solvents can make solar cells safer, cheaper, and more efficient.